How can you experiment with new technologies in a safe environment? How do you know what data you are missing? How can you scale up an app? How do you perform a survey to gain real insights?
Four presentations, four key questions, on our first Open Tea of 2018. Together with The Spindle and Partos, we kicked off a new series of meet-ups to share experiences and challenges in ‘data-informed decision making in development cooperation’.
The use of high quality information and data is an important building block for transparency and effectiveness of development cooperation and humanitarian aid.
The Open Tea format is simple: the host organisation invites a few people to give a short presentation and provide room for questions and exchange. Afterwards, the participants enjoy some drinks while networking and continuing the conversation. It’s a great format to learn about each other’s cases and experiences, and to find new connections and sources.
On January 9th, around 30 people gathered at the Partos office, to hear four stories (find the available slides here).
HumanityX’s innovation environment
Josje Spierings talked about her work with HumanityX, part of the Centre for Innovation of Leiden University. Their mission is “To enable organisations working in the peace, justice and humanitarian sector to spearhead innovations in order to increase their impact on society”. An key aspect is to provide a safe environment to experiment, with attention to “responsible innovation”, including privacy, safety and security in the design.
This is shown in the work on a Humanitarian Chatbot (together with Free Press Unlimited), on using data from electronic vouchers for cash-based support (with the World Food Programme) and on monitoring root causes in conflicts (where HumanityX is a data broker keeping an eye on risks in using or combining data sources).
Cordaid’s performance-based financing and data
Maarten Oranje showed the work of Cordaid in health systems strengthening, using PBF (performance-based financing). This requires good quality data at the source, on a range of quantitative indicators: a health facility enters into a contract to get paid based on numbers of verified cases, supplemented with evaluations and patient satisfaction surveys. Furthermore, payments are only done when the reported data is accurate enough. This leads to a rapid improvement of data quality, and eventually also to a lower burden of verification.
It provides a pretty good picture of what happens in the health facilities, but it fails to answer a major question: How many people are we not reaching, how many unregistered cases are there?
How can we get data to answer that question? Several possible sources of data and possible outreach channels were mentioned, namely the Exclusion Radar of the Leave No One Behind campaign, Red Cross’ Missing Maps, Unicef’s Ureport and BBC’s Action.
Patrick Guyer shared Oxfam’s Datalab ambitions, and their challenges in developing a “prototyping way of working” with a focus on creating 2-way streets of data. The data provided by citizens should empower those citizens themselves as well, and not only serve as input for the programs and campaigns of the organisation.
In Uganda, they developed the CashTime app, the result of a hackathon followed up by further development. The app aims to help users manage their household incomes and spendings, and to set up saving goals, and includes budget overviews and financial advice.
At the same time, the aggregate data of all users helps Oxfam get real-time insight in how successful people are in achieving their goals, and how much of their spending goes towards food, healthcare or education. It also opens the way to research potential indicators to predict success. Of course the data only has limited use. It is only representative for the users of the app, not for the population as a whole.
Discussion followed on possible ways to scale the app, lower costs and reach more users, to make the data more representative. An alternative direction could be to use such an app specifically to monitor your own programmes, and not have the ambition to reach users outside of the programme’s target groups.
RNW Media and surveys
Bec Connelly of RNW Media shared her experiences with doing surveys in a corporate environment, and the many aspects that influence the data coming out of it. Surveys are often used just ‘because we can’, and often with insufficient knowledge about the design and bias subtleties. If you do a survey straight after an event, you’ll mostly get information about the experience of participants. If you wait longer, you’ll be able to get more of a reflection on the event. If you ask for reviews, you’ll mostly get reviews from people who either had an particularly good or bad experience.
This highlights a broader challenge around becoming ‘data-driven’: you need to understand the shortcomings and strengths, and to convince your organisation to invest in this. The afternoon ended in a discussion of a discussing the general question “How can I make my organisation more data-driven?”
To help answer that question, Data4Development and The Spindle aim to develop a community of data professionals in the development sector, to exchange experiences, share examples, and keep each other motivated and equipped with approaches and arguments to show how a data-informed approach can improve the cost-effectiveness and impact of our work.
More events to follow
To stay up to date about our next meetings, join our Meetup group. We’re planning Open Teas in May, September and November, as well as an IATI event in early March, and will participate in the Partos Innovation Festival next Autumn.
If you like to host an Open Tea at your organisation, or if you like to present your case and discuss your questions with a group of peer data professionals, let us know, via the group or by contacting Maaike Blom at [email protected] or Anne-Mare Heemskerk at [email protected]